[MXML-3-05] Linear Regression [5/7] - Locally Weighted Regression (LWR) скачать в хорошем качестве

[MXML-3-05] Linear Regression [5/7] - Locally Weighted Regression (LWR)

*** Dubbing: [ English ] [ 한국어 ] In this video, we will look at the Locally Weighted Regression, LWR. We will take an overview of LWR and create a weighted cost function for LWR. Then, let's implement LWR using the scipy and scikit-learn libraries and compare the two results. And let's learn Boston's housing price data using LWR. For nonlinear regression, we need to choose an appropriate nonlinear function and fit that function to the data. You can use nonlinear functions such as polynomial, sine, logarithmic functions, etc. However, it is difficult to choose an appropriate nonlinear function in advance. It's hard to know whether it's better to choose a polynomial function, a sine or logarithmic function, or something else. In multidimensional space, it's even harder to know. This is called parametric methods. LWR is a type of non-parametric method that performs nonlinear regression without choosing a nonlinear function. LWR is a non-parametric method capable of nonlinear regression, but it takes a long time to predict because a linear regression must be performed for each test data point. LWR is a lazy learner like KNN. #LinearRegression #LocallyWeightedRegression #LWR